In order to reach the project goals, the following steps were performed:
Context analysis. It is necessary to start the project with an analysis of the existing datasets, cloud software and available solutions for network data analysis. Due to the reason, that ENQT is located in Germany, the main focus was on the technologies available in the country. It was found that there is no official base stations database in Germany, which can be directly used in the project to identify locations of all currently operating base radio stations. The whole German mobile network infrastructure is under control of the three mobile operators. They provide their own independent network coverage maps, which are usually not very accurate, and also keep information about base stations in a secret. Luckily few open data projects exist, which provide cell tower data and which can be used in our project.
After the context analysis it was decided to create our own base stations database, which can combine data from different resources. The processing is implemented with Python programming language and the database is stored as SQL database. It is daily updated, that allows to always track the base stations which are currently in operation. After analysis of the database it was defined, that Telefonica has the biggest amount of operating base stations (approx. 100 000 LTE stations and 120 000 GSM stations) in Germany. Although theoretically it gives an advantage over the competitors, a lot of stations are located quite close to each other especially in urban areas and can form clusters. Therefore, the database itself is not enough to make any final conclusions and network coverage map is required.
Analytical model. Using the recent publications about coverage analysis for cellular networks and information from the context analysis step, an analytical model was created, which uses data from the base station database. The programming code was implemented in Python, which has a big number of convenient tools for data analysis. The created analytical model is able to plot signal strength maps with different metrics for the main mobile operators and network technologies (GSM, LTE) currently available in Germany.
Testing. Within this step testing of the model on a test server and adjustment according to customers’ requirements were conducted. User can easily change the input parameters, such as operator, network technology and metrics to plot. The interactive map shows all the base stations available and allows to check signal strength at different locations by the clear red-yellow-green colour scheme. The first results look very promising. The general mobile networks patterns in Germany were defined, e.g. GSM coverage area is much bigger and more uniform comparing to the LTE, where base stations are mostly located in urban areas and along the main highways.
Measurement data. In order to increase the map building quality, network signal measurement data obtained by ENQT devices were also used. The measurement points were added to the analytical model to improve the accuracy of the interpolation. However, due to relatively small amount of measurements and the reason, that most of the measurements are conducted in city center, currently there is no significant improvement in the coverage map accuracy. It is planned to use the measurements in the next versions of the code.
Integration. As soon as the testing was conducted and the major corrections were applied, integration of the current version into the web-based information platform of ENQT was done. Both base station database and coverage map are currently available for the customers and together with the other online services provide a great analytical experience, which has never been accessible before.
